Research On Image Compression Coding Using PCNN

The image compression technique is a way of utilizing the correlation of the image between the adjacent pixels, as far as possible to compress the redundancy between the pixels, thereby reducing the transmission bandwidth and storage space. However, the main ambition of the traditional image compression technology which rarely take into account the human visual redundancy is to reduce the coding redundancy and redundant pixels. This paper presents a method using Pulse Coupled Neural Networks (PCNN’s) for image compression based on human visual characteristics. The main studies of this paper are as follows:(1)We used an automatic parameter setting method of PCNN for image quantization. This method successfully determines all the adjustable parameters in PCNN and does not need any training and trials as required in previous studies. In order to obtain better quantitative results, we make some appropriate improvements for the method above.(2)With the consideration of the PCNN parameters、the quantizer levels and the similarity between PCNN and human visual, we choose PCNN for image quantization processing. The results accord with the human visual characteristics.(3)Combination with Huffman coding and Run-length coding, we get a significant enhancement and an effective compression for our experimental image. In the end, the algorithm of Segmented Image Coding is described simply, and the three algorithms is discussed distinctly.The experimental results demonstrate that the algorithm based on human visual characteristics and PCNN not only has high compression efficiency, but also highlights the details of the portion of the image. This method has a broad application prospects in image compression.